In the second quarter of 2026, Russian banks' claims on companies grew by 3.8 trillion roubles, or 3.7%, after growth of 0.7 trillion and 0.7% in the first quarter — a more than fivefold acceleration in three months. In its quarterly banking sector review, published on 2 September, the Bank of Russia gives a reason for that acceleration that its own press release the same day does not mention: the budget spent less.
The sentence
“Growth rates accelerated, among other things, because of lower budget spending — 11.5 trillion roubles in Q2 2026 (1.4 trillion roubles lower than in Q1 2026): companies raised financing in anticipation of forthcoming payments under state contracts.”
That is the central bank describing a mechanism, not an analyst inferring one. The state pays later; the contractor goes to a bank; the bank lends against a budget payment that has not arrived yet. The review attributes the spending figure to Minfin's data. The press release issued the same day, “Lending growth accelerated in all segments”, carries the acceleration but not the explanation.
What drove the acceleration
Rouble corporate loans accounted for most of the growth, rising by 3.4 trillion roubles or 4.0%. A large share went to construction companies — 1.2 trillion, of which 0.6 trillion went to housing developers — and to industrial companies, which received 1.0 trillion. Developers borrow at below-average rates: 9.9%, against 14.7% across the rouble corporate portfolio. The review attributes the difference to funds accumulating in escrow accounts. Small and medium-sized business debt rose more modestly, by 1.8% or 0.1 trillion, while holdings of corporate bonds increased by 0.5 trillion. The CBR kept its 2026 forecast for growth in claims on companies at 7–11%.
The obligation does not disappear
Add the review's own two numbers and the first quarter's spending was 12.9 trillion roubles. State spending falls by 1.4 trillion in a quarterly statistic, and the obligation behind it does not go anywhere — it waits.
The same document says when the money is expected to arrive. Explaining why it raised its forecast for client funds at banks, the CBR writes:
“We raised the initial range by 2 percentage points, because we expect a substantial increase in client funds in H2 2026: at the end of the year companies traditionally receive high payments under state contracts, while in December household funds will grow on bonuses and the advance payment of January transfers.”
So one document gives both sides: companies borrow in the second quarter because the payment has not come, and the payment comes at the end of the year. This is a deferral, not a saving — the money is delayed, not cancelled. The cost of the deferral is the interest the contractor pays at the corporate rate while it waits; the weighted average on the rouble corporate portfolio is 14.7%.
It also explains a pattern visible from outside: a December surge in budget spending, and unpaid bills that can grow while budget execution looks orderly.
What it is doing to loan quality
The review is candid about the other side. The share of problem loans to legal entities, including SMEs, rose 0.2 points to 11.8%, and the volume of problem debt grew by 0.5 trillion roubles, or 4.5% after adjustment for currency revaluation. It attributes the increase to risky restructurings at individual industrial companies, 0.2 trillion, and to a worsening financial position among some borrowers in real estate and the oil and gas industry, 0.1 trillion each.
In the SME portfolio the problem share rose 0.9 points to 20.0%. That is a fifth of that loan book measured in roubles, not one borrower in five. The review gives a reason for its own regulatory response, and the wording is worth reading twice: so that banks would help SME borrowers “affected by attacks on warehouse, production and trading premises”, the regulator recommended that they restructure those loans. The recommendation is its information letter of 7 August 2026, No. IN-03-55/27.
Provisions and good collateral cover about 50% of problem loans, and about 70% in the SME segment. The uncovered part is 5.8 trillion roubles. The CBR puts this at up to about 60% of the capital buffer above the regulatory ratios and judges it manageable because the risks will materialise gradually.
Profit, meanwhile, is high. The sector earned 2.3 trillion roubles in the first half, 36% more than a year earlier, and 1.1 trillion in the second quarter. The CBR raised its 2026 profit forecast to 3.9–4.4 trillion from 3.4–3.9 trillion.
The same pattern, at a much smaller scale
Rosstat's January–July report, published the same day, contains a line that points the same way on a scale several orders of magnitude smaller. Total overdue wage arrears at the end of July were 2,039 million roubles. Of that, 34.5 million — 1.7% — was owed because money did not arrive from budgets on time. Within that fraction:
“Of the total amount of arrears due to budget underfunding, 82.5% fell on the federal budget, 17.5% on local budgets, and there were no arrears from the budgets of the constituent entities of the Russian Federation.”
The sums are trivial. 34.5 million roubles is a rounding error beside the 11.5 trillion above, and the entire arrears figure is under 1% of one month's wage bill. This measures nothing on its own. What it does show is where the delay lies: when the state is late paying wages, it is the federal budget and not the regions. And the line moves fast when cash stumbles — in April 2026 it was 787.6 million, 62 times the March figure, then 618.5 million in May and 25.9 million in June.
What this changes in our model
No indicator moves today; the review publishes no series our parser loads. It bears on three of them, in different directions.
The deficit scores 65 out of 100 — a rolling twelve-month deficit of 3.3% of GDP through July. If spending is being pushed toward December, the current reading is flattered relative to where the same measure will sit once December is in it.
Overdue receivables is the channel where this would show, and it currently carries no score at all: its latest reading has aged out of the window the model allows, so the site says it does not know. It counts overdue bills between organisations, not sums the state itself owes, so a contractor waiting on a budget payment enters the series only once it stops paying its own suppliers. Even that second-hand trace is dark in the quarter the central bank describes the first-hand one building up.
Money supply scores 85. Credit to companies of 3.8 trillion roubles in a quarter is money created by the banking system, and in the second quarter it, rather than the budget, was doing the creating.
What we do not know
The review does not say which state contracts or which industries borrowed against expected payments; “1.0 trillion to industrial companies” covers all of manufacturing. It gives no split of the 11.5 trillion between the federal budget and the wider budget system. And it does not explain why budget spending fell by 1.4 trillion — deliberate stretching, slow contracting and delays on the contractors' side would all produce the same line, and the review does not distinguish among them.
The information letter of 7 August is cited but not reproduced, and the review gives no count, no volume and no regional breakdown for the SME borrowers whose premises were attacked. Rosstat's arrears figures come from organisations' own reporting, exclude small businesses, and a company that files nothing appears nowhere in them.